Minimum-variance strategies – which aim to reduce swings in portfolio prices and typically consider both share-price volatility and intra-stock correlation – have gained much traction since the global financial crisis.
Investors have favored them not just as a risk-reduction tool, but also because low-volatility portfolios have typically been shown to outperform. This latter phenomenon has been referred to as the ‘low-volatility anomaly,’ as it contradicts a traditional premise of portfolio theory that explains asset returns as a compensation for risk.
But are minimum-variance concepts guaranteed to beat benchmarks in all markets? Do such portfolios carry additional and unintended risks that investors should be mindful of?
A new research paper1 by Anand Venkataraman, Head of Product Management at STOXX, attempts to answer these questions by studying the relative performance of minimum-variance strategies and employing a factor-based optimization approach.
Unveiling the extent of the low-volatility premium
The author considered the performance of a global, market-capitalization-weighted universe of 1,800 stocks, which acts as benchmark; and of two derived minimum-variance portfolios: one constrained to the benchmark’s characteristics such as country/industry allocations and style exposures, and one unconstrained. The period analyzed runs from October 2002 to March 2019.
The first finding reveals that, relative to the benchmark, minimum-variance strategies not only reduce volatility persistently, but enable investors to harvest significant low-volatility premia. This is true even after controlling for style, industry and country factors.
While the benchmark generated a total return of 324%2 over the entire period, the unconstrained minimum-variance portfolio returned 439% and the constrained minimum-variance portfolio yielded 504%.
Breaking down the factor drivers of a minimum-variance portfolio
However, this outperformance comes at a price. Selecting and weighting stocks in order to reduce portfolio risk leads to exposures to other style factors with potential negative performance contributions.
Venkataraman breaks down the strategies’ performance into their underlying systematic drivers to understand low volatility’s relevance on returns in comparison to other factors that arise unintentionally as a result of holding low-volatility stocks — such as momentum, liquidity or value.
The unconstrained minimum-variance portfolio outperformed its benchmark in the period analyzed by 115 percentage points. That active return can be decomposed into a specific return and a factor-related contribution, Venkataraman explained. The factor component, in turn, is further broken down into style, country, industry, currency and market factors. With 93 percentage points, the specific component dominates the active return while factor contribution accounts for 22 percentage points.
Within the latter, the style factor has a surprisingly low effect — only 7 percentage points. A further decomposition of the style factor reveals that the low-volatility sub-factor contributed a noticeable 169 percentage points. However, negative contributions from the leverage, liquidity, momentum and value style sub-factors more than offset the positive effect of low volatility. Size, exchange-rate sensitivity and growth were the other sub-factors that impacted positively.
The research run a similar study on emerging markets, as well as on some regional and country markets, with similar results.
Paying a price to contain negative factor effects
Being aware of such interactions among factors, investors may consider applying appropriate counter-measures. Constraining unintended style exposures may help reduce their significance and negative performance contribution. This, again, doesn’t come free: it causes a lower reduction in portfolio risk and a decrease in the low-volatility premium,Venkataraman said.
The study provides a useful appraisal of the performance and style exposures of implementing minimum-variance strategies and the effect of controlling for industry and country factors. Additionally, it unearths the price of constraining the exposure deviations of such minimum-variance portfolios from their respective market-capitalization benchmarks.
1 Venkataraman, A., “The Low Volatility Premium – An Analysis of Factor Exposures of Minimum-Variance Strategies,” STOXX, May 1. 2018.
2 Gross returns in dollars.